Linear Equalizer Control Loops for Transparent Link Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Automatic Gain Control (AGC) circuits are non-transparent to link training in linear signal conditioning devices, leading to suboptimal bit error rate, power management, and electromagnetic interference issues, and can distort transmitter waveforms, making it difficult for receivers to recover data correctly.
Innovation Solution
A linear equalizer system with a controller comprising a gain control loop, a crossbar loop, and a driver loop, each with a replica of the equalization paths, allowing for transparent link training by comparing input and output levels to maintain channel gain, using nested AGC loops to control gain variations independently of configuration and minimize noise amplification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional AGC circuit is used to control gain, then output signal level is maintained, but link training transparency is lost and bit error rate performance deteriorates
Solution Approach 1:
The equalizer is divided into multiple independent equalization paths (first path with first VGA, second path with second VGA, etc.), each capable of being independently controlled during link training. This segmentation allows the system to maintain transparency to link training while still providing gain control when needed, resolving the contradiction between output level control and link training transparency.
Solution Approach 2:
The system dynamically switches between different operational modes: during link training, multiple equalization paths are activated to maintain transparency; during normal operation, gain control is applied to maintain output signal level. This dynamic adaptation allows the system to optimize performance for different operational phases, improving both link training transparency and bit error rate performance.
2Stability of the object's composition
If AGC circuit adjusts gain to maintain output level, then signal level stability is improved, but link training configuration optimization is prevented
Solution Approach 1:
The system performs preliminary gain control configuration during link training by activating multiple equalization paths, allowing the link training to complete its configuration optimization. After link training is complete, the system then applies AGC gain control to maintain output signal level stability. This preliminary action ensures that adaptability is maintained during configuration while stability is achieved during operation.
Solution Approach 2:
The system changes operational parameters (number of active equalization paths, gain control activation) based on the operational phase. During link training, parameters are set to maximize adaptability (multiple paths active, gain control disabled); during normal operation, parameters are adjusted to maximize stability (fewer paths needed, gain control active). This parameter adaptation resolves the contradiction between stability and adaptability.
3Power
If conventional AGC is used in linear equalizer, then gain control is achieved, but electromagnetic interference performance deteriorates
Solution Approach 1:
The system applies gain control locally to specific equalization paths rather than uniformly to all paths. During link training, multiple paths operate without aggressive gain control, reducing EMI. During normal operation, gain control is applied selectively to maintain output level while minimizing EMI impact on other paths. This localized approach allows gain control capability while reducing overall EMI.
4Measurement precision
If limiting devices are used for gain control, then output level is maintained, but transmitter waveform is distorted and data recovery becomes difficult
Solution Approach 1:
The system uses replica equalization paths that copy the structure and characteristics of the main equalization paths. These replica paths are used during link training to maintain waveform integrity while providing gain control functionality. By using copies rather than direct limiting devices on the main paths, the system maintains waveform integrity while still achieving output level control capability.
Data Source
AI summary
An apparatus for equalizing channels is provided, which is generally transparent to link training. The apparatus generally includes equalization paths formed by an input circuit, a crossbar, and an output circuit and a controller. Each equalization path is coupled to at least one of the channels, and a controller has a VGA loop, a crossbar loop, and a driver loop. The AGC loop receives a first reference voltage and provides a gain control signal to the input circuit, and the gain control network comprises a replica of at least one of the equalization paths. The crossbar loop receives a second reference voltage and provides a crossbar control signal to the crossbar. The driver loop receives a third reference voltage and provides a driver control signal for the output circuit.